Mathematician Claims Riemann Hypothesis Proof With AI, 250,000-Page Lean Check Begins

A mathematician at Anthropic claims to have proven the 167-year-old Riemann hypothesis using the AI model Claude, producing a Lean proof spanning more than 250,000 pages that independent experts have begun verifying on Friday.

Alpöge, 34, a researcher at the American AI company Anthropic, wrote that he believes he has found a proof of the Riemann hypothesis in collaboration with Claude, according to De Telegraaf.

The hypothesis, formulated by Bernhard Riemann in 1859, concerns the distribution of prime numbers — those divisible only by one and themselves, such as 2, 3, 5, 7 and 11 — and asserts that the non-trivial zeros of the zeta function all lie on a single critical line.

If confirmed, the proof would also establish the Generalized Riemann Hypothesis, a far broader conjecture with deep consequences for number theory and cryptography. Alpöge published the argument in Lean, a proof-assistant language that allows computers to check every logical step. The file exceeds 250,000 printed pages; a natural-language version would be even longer, he noted.

“Claude danced, sparkled,” Alpöge wrote on his site. “But just before the finish line, Claude left me behind.”

The Clay Mathematics Institute included the Riemann hypothesis among its seven Millennium Prize Problems in 2000, offering $1 million for a solution accepted by the mathematical community after two years of scrutiny in a refereed journal, according to Ad.

Andrew Sutherland, a senior research scientist at MIT, told Investing.com that he successfully verified the submitted Lean code on his computer. Stanford Ravi Vakil added, hours after Alpöge’s post, that a postdoctoral researcher in his department had also submitted an independent proof of the hypothesis.

Terence Tao, a Fields Medalist at UCLA, cautioned that the sheer length of the code means full peer review will take considerable time. “It is going to take quite a while before experts have checked everything and signed off,” he said.

Buckmaster and Alpöge released a new Navier–Stokes proof Friday alongside the Riemann announcement, stating it avoids issues identified in the OpenAI attempt. “After last time we couldn’t wait,” Buckmaster wrote.

Mathematician Claims Riemann Hypothesis Proof With AI, 250,000-Page Lean Check Begins
Photo: Periskopi
Problem Proposed Solver AI System Status
Riemann Hypothesis Levent Alpöge (Anthropic) Claude (internal version) Lean code published; independent verification started
Navier–Stokes Existence & Smoothness Alpöge & Tristan Buckmaster OpenAI model (Sept); own methods (Nov) New proof posted Friday; earlier OpenAI claim disputed
Prize Clay Mathematics Institute $1 million each Requires journal publication + 2-year acceptance

Alpöge acknowledged he does not yet fully grasp all the mathematical concepts Claude generated.

Born on Long Island in 1992 to Turkish-American parents, Alpöge was a finalist in the 2010 Intel Science Talent Search; an asteroid, 25898 Alpöge, was later named in his honor.

He graduated summa cum laude from Harvard in 2014, spent a year at Cambridge as a Churchill Scholar, and earned his Ph.D. at Princeton in 2020 under Fields Medalist Manjul Bhargava. Before joining Anthropic, he held an NSF postdoctoral fellowship at Columbia and Harvard and served as a Junior Fellow in the Harvard Society of Fellows from 2021 to 2025.

The proof remains unpublished in a peer-reviewed journal. The Clay Institute’s two-year clock starts only after such publication. Until then, the $1 million prize — and the cryptographic implications of a confirmed Riemann hypothesis — stay in limbo.

Photo of author

World Desk

World Desk at Archyde.

NHS Leaders Urge Over-40s to Get Free Checks as High BP Admissions Rise 77% in Six Years